Multiple Linear Regression Modeling in Determining the Contribution of Landshape Factors on the Quantitative Attributes and Diversity of Trees

Introduction : Mountainous forests are a major part of the northwestern forests of Iran that are tasked to protect biodiversity. Therefore, mountainous conditions create diversity in landforms. Multiple linear regression models are used to create ‎the relationships between different attributes of the forest stands and various landforms to assess other characteristics ‎indirectly‎. In protective forests, the basal area and species diversity are the main attributes that are considered significant factors in forest planning and management. The main objective of current study was to understand the association of basal area and tree species diversity index with DBH, tree height, canopy cover, and landform indices in a natural mixed-species protected forest in northwest of Iran and develop estimation equations by using a multiple linear regression model. Methods The data collection from direct measurements of the stand to obtain the basic data (tree characteristics used to estimate the volume, basal area, etc.) was the systematic random method in various topographic conditions with diverse slopes and aspects located in Arasbaran forest, in the east Azarbayjan province of northwest Iran. After the data collection, in the second stage of the study, the dependent (basal area and diversity index) and independent (forest attributes and landform indices) variables were determined and the relationship between the factors was evaluated with the help of the multiple linear regression statistical method. Results The results of multiple linear regression showed that the Shannon diversity index was influenced by tree height, tree average crown diameter, AspE, TRASP, and SEI land shape indices. Cohen's f ² effect size of factors was 0.252, which is in the medium effect size range. In addition, we found that the basal area of trees was influenced by tree dbh, crown area, AspE, and SEI land shape indices. Cohen's f ² effect size was in the strong effect size category (0.323). In general, the multiple linear regression model results in the present study showed that the independent variables had a significant effect on dependent variables and these effects were at an acceptable level in the most cases. Discussion Our study highlights that modifications in basal area and diversity were corrected with landform indices which can be used as a base for organizing forest management plans including basal area increasing in protects forests with low commercial volume objects.